Follow-up: link thoughts and learnings through explicit associations #32

New Issue

2026-04-21T22:27:54Z

sgcommand commented

2026-04-21 22:27:54 +00:00

Summary

Add follow-up support so thoughts and learnings can be explicitly linked and associated rather than living as isolated records.

Why

Structured learnings are useful on their own, but they become much more practical when they can point back to the raw thought process, source notes, or working memory that produced them. Likewise, thoughts should be able to reference the curated learning that came out of them.

This keeps the distinction clear:

thoughts = raw notes, working memory, intermediate observations
learnings = curated, durable lessons

But it also avoids silos between the two.

Recommended scope

Add explicit association support between thoughts and learnings, for example:

a learning may reference one or more source thoughts
a thought may reference zero or more related learnings
links should be queryable in both directions
associations should be lightweight and not require workflow automation

Suggested model options

Possible implementations could include one of:

direct foreign-key style fields for simple one-to-many linking
a join/association table for many-to-many relationships
typed relationships if future expansion to other entities is expected

Acceptance criteria

A learning can be linked to one or more thoughts
A thought can expose related learnings
Links are retrievable in both directions
The model preserves the distinction between raw thoughts and curated learnings
The design does not force heavy workflow or auto-promotion behavior

Non-goals

automatic promotion of thoughts into learnings
semantic memory graph or advanced associative inference
replacing either thoughts or learnings with a single merged model

Rationale

This gives AMCS a clean bridge between working memory and curated knowledge without turning v1 into a giant knowledge-graph project.

## Summary Add follow-up support so thoughts and learnings can be explicitly linked and associated rather than living as isolated records. ## Why Structured learnings are useful on their own, but they become much more practical when they can point back to the raw thought process, source notes, or working memory that produced them. Likewise, thoughts should be able to reference the curated learning that came out of them. This keeps the distinction clear: - thoughts = raw notes, working memory, intermediate observations - learnings = curated, durable lessons But it also avoids silos between the two. ## Recommended scope Add explicit association support between thoughts and learnings, for example: - a learning may reference one or more source thoughts - a thought may reference zero or more related learnings - links should be queryable in both directions - associations should be lightweight and not require workflow automation ## Suggested model options Possible implementations could include one of: - direct foreign-key style fields for simple one-to-many linking - a join/association table for many-to-many relationships - typed relationships if future expansion to other entities is expected ## Acceptance criteria - [ ] A learning can be linked to one or more thoughts - [ ] A thought can expose related learnings - [ ] Links are retrievable in both directions - [ ] The model preserves the distinction between raw thoughts and curated learnings - [ ] The design does not force heavy workflow or auto-promotion behavior ## Non-goals - automatic promotion of thoughts into learnings - semantic memory graph or advanced associative inference - replacing either thoughts or learnings with a single merged model ## Rationale This gives AMCS a clean bridge between working memory and curated knowledge without turning v1 into a giant knowledge-graph project.

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: wdevs/amcs#32